Reef Fish

Background

This is an investigation into a survey dataset generated by the National Oceanic and Atmospheric Administration’s Southeast Fisheries Science Center (NOAA Fisheries) and the National Centers for Coastal Ocean Science (NCCOS). This dataset includes coral reef fish survey data from three distinct regions along the Florida reef tract: 1) Dry Tortugas, 2) Florida Keys from Key West to Miami, and 3) Miami to Martin County. The data were collected using a two-stage, stratified random survey design, employing the stationary point count method (7.5m radius cylinder) as part of the ongoing National Coral Reef Monitoring Program (NCRMP).
Particular data spans from May 2014 through 2018 although the survey is still ongoing. Over 250,000 data points capture fish species, depth, length, and counts at more than 5,000 individual locations across the Florida Reef Tract (see map of all geopoints in image to the right). The explicit purpose of the data and the countless hours devoted into such an extensive survey are to "To provide continued reef fish and habitat monitoring in Florida’s coral reef tract; to assess population and habitat trends, fish-habitat associations; and to assess ecosystem responses to natural events (e.g., hurricanes), management measures and anthropogenic impacts".

This data is particularly interesting to me as I have spent a fair bit of time underwater in the Florida reef tract. Fishing and diving in the region is some of the best in the contiguous United States, and it is a privelege to have resources devoted to researching and protecting this environment. While I am most familiar with the region of reef east of Biscayne Bay, the entire dataset is intriguing. I want to determine how similar my observations are to those of seasoned surveyors and see if I can gain any insight into species distribution, habitat, or patterns that will help me be a more informed diver.

Data is structured like that in the table below. In the table below, there were three gray triggerish seen at 23.5 and 23.9 m deep. Two were 29 cm in length and one was 22 cm. There were also 36 tomtates of either 10 or 12 cm. To better quantify overall observations, any fish with multiple individuals of one length recorded at a single location and timestamp were "flattened" into multiple equivalent rows each of one observation at that length.

Time Longitude Latitude Depth Species # Scientific Name Common Name Length Number Seen
2014-05-06T00:00:00Z 27.17738 -80.04002 23.5 16 Balistes capriscus gray triggerfish 22 1
2014-05-06T00:00:00Z 27.17738 -80.04002 23.9 16 Balistes capriscus gray triggerfish 29 2
2014-05-06T00:00:00Z 27.17738 -80.04002 23.9 90 Haemulon aurolineatum tomtate 10 6
2014-05-06T00:00:00Z 27.17738 -80.04002 23.5 90 Haemulon aurolineatum tomtate 12 30

Fish Diversity

One of the first things we might want to investigate in this sampling study is the species diversity. Diversity can be quantified a number of ways, but the easiest is looking at the quantity of unique species observed and their relaitve prevalence. In our selected data set from 2014 to 2018 in the Florida Reef Tract, 276 different common names were recorded. Most of these represent independent species while a few of them are unidentified subtypes like "surgeonfish species". The most observed species is the masked goby with over 236,000 noted individuals. Several species were only observed once - among them were: [yellowfin group, tiger goby, papillose blenny, lookdown, dwarf goatfish, etc.].
To visualize this species diversity, two treemaps are presented below. The first treemap includes all species observed more than 250 times while the second is all species observed less frequently. The treemaps are separated to not obscure the rarer fish. Each tree map can be interacted with by selecting a group name to enlarge or by hovering over the cell to see summary stats like the total observations, average length, and average depth. The maps are shaded by average depth. Click the top by the name to return to the parent group.

Common Fish
Rare Fish

These maps are information dense and require a little bit of time to process. Some notable observations are:

  • The 176 rarer species account for only 0.76% of all observations despite being 64% of total diversity.
  • The sardine group contains both the shallowest average species depth (scaled sardine, 2.38 meters) and the deepest (redear sardine, 25.66 meters).
  • The Goby group has the greatest representation with 17 species noted (+1 general Goby). The next highest is Parrotfish with 15 and Snapper with 12.
Somewhat humorous to me, there are 46,000 observations noted as "grunt species" - by far the greatest section of the grunt group. It is comforting to know that talented fish surveyors are just as sure about grunt classification as I am.
There are also a number of never-recorded species that feel underrepresented based on my own experiences. A few that come to mind are the cobia and the cubera snapper.

Community

This is an investigation into a survey dataset generated by the National Oceanic and Atmospheric Administration’s Southeast Fisheries Science Center (NOAA Fisheries) and the National Centers for Coastal Ocean Science (NCCOS). This dataset includes coral reef fish survey data from three distinct regions along the Florida reef tract: 1) Dry Tortugas, 2) Florida Keys from Key West to Miami, and 3) Miami to Martin County. The data were collected using a two-stage, stratified random survey design, employing the stationary point count method (7.5m radius cylinder) as part of the ongoing National Coral Reef Monitoring Program (NCRMP).

Snappa

Add more context, segue into next conversation

Describe figure

Describe first image.

Supplemental Fun

One of the more challenging components from Wordle's design is that duplicate letters are sometimes difficult to identify as a single green or yellow square provides no indication about potential repeats. Additionally, guessing a duplicate letter does not maximize the amount of information provided in each guess. Therefore, words with fewer than 5 characters are less conventional and certainly less than optimal guesses. Did the author do anything to make these words more or less likely than a more complete dictionary?

Additional Data Exercises

  • Exercise 1
  • Exercise 2

Links